Am 30.04.2012 20:50, schrieb Tristan Matthews: > ----- Original Message ----- >> From: "Werner Dittmann" <[email protected]> >> To: [email protected] >> Sent: Monday, April 30, 2012 3:11:37 AM >> Subject: Re: [Ccrtp-devel] CryptoContext: uninitialized value >> >> Initialization to NULL is correct. In case of non-authentication mode >> this could >> lead to problems if macCtx contains some random value and delete uses >> this value >> to free the macCtx. > > In that case, could the attached patch be applied? > > Best, > Tristan >
In parts it's already included, only the "empty" constructor is missing, will be added soon. Thanks, Werner >> >> This problem did not show up because SRTP usually uses an >> authentication mode and >> macCtx was initialized eventually. Note: SRTP shall always use an >> authentication. >> >> Regards, >> Werner >> >> Am 30.04.2012 03:45, schrieb Tristan Matthews: >>> [Apologies for cross-posting...accidentally mailed this to >>> bug-commoncpp] >>> >>> I was just wondering if macCtx in CryptoCtx is deliberately _not_ >>> initialized in either of CryptoContext's constructors. >>> I ask because this value is used in an if-statement in the >>> destructor for CryptoContext (see CryptoContext.cpp:180). >>> Note that the body of the if-statement checks against initialized >>> values before doing any work, so the current code is harmless. >>> >>> Best, >>> Tristan >>> >> >> >> _______________________________________________ >> Ccrtp-devel mailing list >> [email protected] >> https://lists.gnu.org/mailman/listinfo/ccrtp-devel >> _______________________________________________ Ccrtp-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/ccrtp-devel
